-----------------------------
Please read the FAQ!
<http://java.apache.org/faq/>
-----------------------------
Craig -
With the db connection pooling, did you try to alter the
sqlnet.expire_time (sqlnet.ora) or connect_timeout_listner
(listner.ora) parameters? Just curious.
greg
Craig R. McClanahan wrote:
>
> -----------------------------
> Please read the FAQ!
> <http://java.apache.org/faq/>
> -----------------------------
>
> "Priatel, Mark (ott)" wrote:
>
> >
> > > (Actually, while writing this I just restarted apache to fix the problem,
> > > and it told me httpd wasn't running...hmmm)
> >
> > >We've had this problem with servlets hanging a few times and it was always
> > a
> > >bug in our own servlets/classes that caused it. It usually had to do with
> > bug's
> > >in the code dealing with the pool of DB connections. Are your servlets
> > >connecting to a DB by any chance?
> >
> > As a matter of fact, I am! =)
> >
> > I'm using the DbConnectionBroker class by Marc Mnich. Presently, I only
> > have about 4 servlets running that each make use of their own broker, and
> > since they're not being hit often (except by me and my team), it shouldn't
> > be too hard to track down the source of the bug. If this (-my bug) is only
> > happening after a long period of inactivity, is it possible that something
> > is happening when the servlets are being destroyed ? Because these servlets
> > are just skeleteons, none of them implement destroy() at all...
> >
> > Anyhow, if you can think of any good places to start looking, let me
> > know...I'll start checking for loose connections, and destroying the broker
> > on destroy()
> >
>
> This kind of issue (stale connections) bit me as well, although with
> SQLException errors instead of hangs.
>
> I'm using a (home grown) connection pool to talk to Oracle. The initial design
> had the available connections stored in a Last-In First-Out stack, which meant
> that the connection at the "bottom" of the stack was used pretty rarely. It
> seems that Oracle SQL*Net will sometimes disconnect a long-lived connection
> that hasn't been accessed for a significant amount of time, which caused a
> problem for the next user that I eventually allocated that connection to.
>
> Adding logic to periodically check for stale connections, and use the
> connections first-in first-out instead, cleared up these problems.
>
> Craig McClanahan
>
> --
> --------------------------------------------------------------
> Please read the FAQ! <http://java.apache.org/faq/>
> To subscribe: [EMAIL PROTECTED]
> To unsubscribe: [EMAIL PROTECTED]
> Archives and Other: <http://java.apache.org/main/mail.html>
> Problems?: [EMAIL PROTECTED]
::: Greg Barish ([EMAIL PROTECTED]) ::::::::::::::::::::::::::::::::::::::::::
--
--------------------------------------------------------------
Please read the FAQ! <http://java.apache.org/faq/>
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Archives and Other: <http://java.apache.org/main/mail.html>
Problems?: [EMAIL PROTECTED]
